IPT 29 2009 11/6/09 09:29 Page 46 Biopharma The Emergence of Antibody Fragments and Derivatives By Gurdeep Singh Antibody fragments have distinct advantages over antibodies, and with Athwal at Novozymes both commercially viable fragments in the marketplace and numerous BioPharma UK candidates in various stages of development, they look set to offer a credible therapeutic alternative to full-length mAbs. The groundbreaking work of Köhler and Milstein, first radioisotopes), and improved diagnostic techniques (for published in 1975 (1), introduced methodologies for the example, rapid clearance of radio-labelled imaging production of monoclonal antibodies (mAbs) of precise agents). The smaller size of antibody fragments may also and consistent specificity, and has for this class of proteins allow interactions with cryptic epitopes, such as enzyme being able to deliver their therapeutic promise. active sites, which mAbs may not be able to access. Currently five mAbs – Herceptin® (trastuzumab, Similarly, removal of the constant (Fc) region can decrease Genentech/Roche), Avastin® (bevacizumab, Genentech/ non-specific effector function and lower potential Roche), Rituxan® (rituximab, Biogen Idec/Genentech/ antigenicity. Such fragments can be viewed as the smallest Roche), Humira® (adalimumab, Cambridge Antibody functional building block from which it is possible to Technology/Abbott) and Remicade® (infliximab, make more complex molecules. Common to all antibody Centocor/J&J) – respectively account for close to three- fragments is a conserved framework region with three quarters of the revenue in this biopharmaceutical sector, complementarity-determining regions (CDRs), which with the total market being made up of just over 20 mAbs. can specifically interact with an antigen. The dominance of the five above-named mAbs is due to a Of the many antibody fragments now described, two number of factors, including: being first to market, market dominant fragments in commercial development are expansion across indications and the inherent accumulation antigen binding fragments (Fabs) and single chain of clinical data, ensuring that newer therapies are used only variable fragments (scFvs). Of the 54 antibody fragments as a second line. However, there are more than 200 mAbs in that have entered clinical studies, 56 per cent are Fabs various phases of clinical trials. Generally, the outlook is very and 35 per cent are scFvs, with the remaining nine per promising for first-generation mAbs, as their exposure to cent being made up of other derivates (2). It should be generic competition is limited. In the main, this can be noted that this distribution can be attributed in part to ascribed to a mixture of technical, intellectual property and the phases of development within the field of antibody regulatory barriers. fragment technologies. With regard to size, scFvs are What is clear is that this sector is commercially attractive, ~30kDa (approximately 85 per cent smaller than but also a very competitive marketplace – both now and for whole mAbs at ~150kDa), and Fabs are ~70kDa, the foreseeable future. These factors go some way to (approximately 65 per cent smaller). More recently, an explaining both the emergence and growth of derivatised even smaller functional unit – called domain antibodies antibody fragment technologies. Antibody fragments also (dAbs) – have been shown to have therapeutic promise; represent part of the natural evolution of the sector, by being these are either equivalent variable heavy or variable light able to utilise the wealth of information already chains from scFvs. Other similar domain antibodies, but accumulated for mAbs, such as humanisation and affinity with less well-known formats, include variable novel maturation. For reference, it is beyond the scope of this antigen receptors (VNAR), derived from cartilaginous fish, review to cover alternative scaffolds such as ankyrin repeats, and variable heavy chains (VHH) from camelids anticalins, adnectins, affibodies and kunitz domain proteins. (Camelidae). Both VNAR and VHH formats possess smaller domain sizes than scFvs (12kDa and 15kDa, ANTIBODY FRAGMENTS respectively), and are similar in size to dAbs. They appear Antibody fragments – by their inherent nature – can have to have diverged from typical murine or human distinct advantages over full-length mAbs. One example immunoglobulins and, as whole antibodies, only have is their ability to penetrate tissues or organs more one variable-like domain. Similarly, unlike conventional comprehensively, thereby providing scope for enhanced murine or human variable domains, the antigen-specific therapeutic effects (for example, the delivery of cytotoxic binding loops (CDRs) are typically longer, allowing 46 Innovations in Pharmaceutical Technology IPT 29 2009 11/6/09 09:29 Page 47 them to access antigen cavities that typical CDR binding Figure 1 loops would be unable to penetrate. Graphical representation of antibody fragments, as derived from a ‘typical’ The general development of phage display isotype G immunoglobin (IgG), a shark Ig new antigen receptor (IgNAR) and technologies has allowed the generation of large libraries a camelid VHH-Ig. Constant domain of fragments and derivatives thereof. These libraries have frameworks for all the antibodies are shown in purple, with the variable been successfully used to screen for the isolation of domains (V) shown in blue, brown, pink and green. Both camel and shark antibody fragments specific for different types of ligands antibodies have a single V domain (VhH and V-NAR), and IgGs have two – ranging from haptens and carbohydrates, through to V domains called heavy (VH) and light (VL) respectively. A selection of the proteins – utilising common high-throughput screening possible antibody fragments which can be derived from these full methodologies. Interestingly, many of these fragments length antibodies are shown, are amenable to protein engineering that is either not with approximate sizes (kDa). possible or difficult to achieve with whole mAbs. For example, the linking of two identical or different scFvs to Figure 2 form a diabody can give rise to either bivalency or Graphical representation of a typical PEGylated Fab, a C-terminal bispecificity – thereby potentially enhancing avidity to monovalent scFv albumin fusion and a bispecific N- and antigens or mediating cross-linking, such as the C-terminally fused scFvs albumin fusion molecule. Approximate sizes recruitment of cytotoxic T-cells to mediate the are shown for the albumin fusions termination of a target cell. Similarly, multimeric forms (kDa). The size of a PEGylated fragment is dependent upon the of other antibody fragments have been described in size and number of covalently linked polyethylene glycol (PEG) the literature, such as Fab2, Fab3, Triabodies and polymers. Tetrabodies (3). Currently, only three FDA-approved antibody production of antibody fragments, both as discrete entities fragments are commercially available: ReoPro® (abciximab, and as fusions, where the antibody fragment is genetically Centocor/Lilly), a platelet aggregation inhibitor Fab; fused to another protein such as albumin. Improvements in Lucentis® (ranibizumab, Genentech/Novartis), an affinity yeast expression have resulted in higher yields and decreased matured Fab fragment derived from Genentech’s anti- or no proteolysis. Similarly, the solubility and stability issues VEGF mAb Avastin® for the treatment of ‘wet’ age-related associated with antibody fragment production have been macular degeneration (vision loss); and more recently, addressed (5); hence, multigram expression and secretion of Cimzia® (certolizumab pegol, UCB) an anti-inflammatory functional antibody fragments and fusions is common. PEGylated Fab for the treatment of Crohn’s disease and This, together with the requirement for inexpensive media rheumatoid arthritis. and ease of scale up to industrial-scale fermenters, bodes well for yeast production to be a viable and cost-effective PRODUCTION ISSUES alternative to mammalian cell culture production. Taking the mAb Remicade® as an example, it is possible to Attempts have been made similarly to express highlight the need to develop cheaper and more efficient antibody fragments using prokaryotic expression methods for the commercial production of mAbs and, systems, such as Escherichia coli (6); however, they can potentially, via antibody fragments. Typically, Remicade® is have inherent limitations. Secretion of antibody administered bi-monthly with a dosing regimen of 5mg/kg; fragments into the culture media can lead to cell lysis and thus, with a conservative estimate of at least half a million associated product loss. Also, prokaryotes do not always patients worldwide being treated, the annual demand have the machinery for correct folding or necessary post- would exceed 1,000kg. This can be compared with other translational modifications associated with eukaryotic therapeutic proteins, where the demand can be 10-fold less protein expression. These limitations can mean that (4). Therefore, production represents one of the main cost- insoluble aggregates are produced. Attempts to re-fold drivers for commercial manufacture of mAbs and, in the such products in vitro can place cost-prohibitive future, antibody fragments. Mammalian Chinese hamster demands on a manufacturing process, as well as the ovary (CHO) cells are the predominant cell lines used in the robustness of a production process at scale. Similarly, the manufacture of mAbs;